Solaris operating system
In the press release jointly issued by Oracle Bone Inscriptions and Sun Company, it is pointed out that Sun Company's Java language and Solaris operating system are "fundamental long-term strategic advantages". This shows the importance of Solaris operating system to Oracle.
As a database supplier, Oracle Bone Inscriptions dreams of owning his own operating system. Without its own operating system, the only choice is to fully support various operating systems on the market and adjust the priority of different operating systems according to market changes. Before 2000, Sun Company was in full swing, and its installed capacity ranked first in many key fields such as finance, telecommunications and energy, so Oracle chose Solaris as the preferred operating system. 200 1 or so, GNU/Linux has excellent performance on the server side, it can run on low-cost x86 processors, and it can also easily download installation files for free from the network. More importantly, a large number of free technicians share their experiences through various channels, and it is easy to find various documents related to GNU/Linux on the Internet. On the contrary, Solaris needs to run on an expensive UltraSparc processor (at that time, the x86 version of Solaris 8 can be downloaded for free, but it is far from robust and practical), and the installation process is complicated, so the system administrator needs special training from Sun Company. With the complete bursting of the internet bubble, enterprises have put forward higher and higher requirements for the cost performance of information systems. In this case, the new installed capacity of Solaris began to decline, and GNU/Linux, which can be deployed immediately without the approval of the financial director, began to occupy the data center. Oracle noticed this trend in time, and launched a customer support plan called Unbreakable Linux in 2002, and began to tilt towards GNU/Linux. It should be noted that "Invincible Linux" is not a new GNU/Linux distribution, but refers to the value-added services such as software update and technical support provided by Oracle to Linux users. At the same time, Oracle Bone Inscriptions began to increase investment in operating system, generously funded several key projects of GNU/Linux community, and tried to build a brand-new GNU/Linux distribution. Due to the complexity of the operating system, Oracle Bone Inscriptions made slow progress on his own operating system. Due to direct competition with IBM's DB2, although the market of Solaris is shrinking, Oracle still takes Solaris as the preferred operating system. In 2006, the installed capacity of GNU/Linux in data center has exceeded 50%, while the installed capacity of Solaris has dropped below 15%. At this time, Larry Ellision, CEO of Oracle Bone Inscriptions, made a wise decision: change the icon of Red Hat to the icon of Oracle Bone Inscriptions, and change the word "Red Hat Enterprise Linux" to "Oracle Bone Inscriptions Enterprise Linux". Oracle enterprise Linux, which is fully compatible with Red Hat enterprise Linux, was launched in one fell swoop, and the customer support plan was launched at half the price of Red Hat. Although there are different opinions in the industry, Oracle finally got rid of its dependence on Solaris by relying on the powerful strength of GNU/Linux community, and was also addicted to its own operating system.
In April, 2006, when asked if he would consider buying Red Hat, Larry Ellision, president of Oracle Bone Inscriptions, said confidently: Oracle Bone Inscriptions will not buy a company that may be eliminated at any time. Nevertheless, Red Hat's successful market share proves that they know the operating system better than Oracle Bone Inscriptions. Oracle can easily make use of the achievements contributed by companies such as Red Hat and Novell to the open source community, but it can't bring the prestige of these companies in the open source field, nor can it bring the loyalty of users to these excellent companies.
This may be one of the reasons why Oracle Bone Inscriptions bought Sun: Sun knows the operating system. As far as file system is concerned, ZFS is the most powerful file system at present; In application development, DTrace can easily find the bottleneck of application from kernel layer and user layer; In terms of system security, Solaris trusted extension has obtained the most comprehensive EAL 4+ certification (including LSPP, CAPP and RBAC). In terms of supercomputing, Texas Advanced Computing Center, currently ranked sixth in the world, runs the Solaris operating system. (It should be noted that the top five supercomputers are all running different versions of GNU/Linux operating system. Despite the slow growth of new installed capacity, there are still about 10% servers running different versions of Solaris operating system in global data centers. Through the acquisition of Sun Company, Oracle Bone Inscriptions can gain the core technology, talents, reputation and existing customers in the operating system field at one stroke.
Therefore, Oracle Bone Inscriptions will not give up the Solaris operating system. The remaining problem is how to continue to develop Solaris operating system. Will Oracle continue to support the current OpenSolaris project, or will it adopt the traditional closed-source approach? My personal opinion is that it is possible for Oracle to republish OpenSolaris project under the framework of GPL license agreement (possibly GPLv3). The biggest problem of OpenSolaris at present is the problem of hardware compatibility. On the x86/x64 platform, there are still a large number of external devices such as sound cards, network cards and graphics cards that are not driven by OpenSolaris. The GNU/Linux community has solved similar problems. The related drivers are basically open source, but because of the incompatibility between GPL protocol and CDDL protocol, OpenSolaris community can't make use of these achievements of GNU/Linux community. I clearly pointed out in the article "Review: OpenSolaris 2008.11"that publishing OpenSolaris project with GPL authorization protocol can solve the dilemma faced by Opensolaris project technically and eliminate the hostility of GNU/Linux community emotionally. It is true that GNU/Linux community can make full use of the highlights of OpenSolaris project, such as ZFS and DTrace, and use GPL licensing protocol, but in the long run, I am afraid OpenSolaris project will benefit more. After all, for an ordinary developer, GNU/Linux without ZFS and DTrace is good enough, but OpenSolaris without sound card and network card driver is not good enough.
Larry Ellision has a famous saying about open source software: "If open source software becomes good and simple enough, we will take it away and use it." Now that the advantages of GNU/Linux in device driver are so obvious, will Larry miss this opportunity?
Java programming language
Speaking of Java programming language, we must first make clear that Java language has three branches: Java SE (Standard Edition), Java EE (Enterprise Edition) and Java ME (Embedded Edition). Java SE is the foundation of Java EE and Java ME, which is similar to the basic research project funded by the National Natural Science Foundation, and basically spends money without making money. Java EE can be regarded as the application of Java language in enterprise-level solutions. Most companies that do Java, except Sun, basically rely on Java EE for profit. In recent years, Java EE has become the de facto standard for enterprise-level solutions. Take China as an example. In 2008, 40% of enterprise applications were based on Java EE and 35% were based on Java EE. Net, 65,438+08% is based on pure HTML and JavaScript. Java ME can be regarded as the application of Java language in embedded devices. For example, most mobile phones now support a subset of Java ME. Before 2006, the manufacturer had to pay a certain license fee to Sun for each mobile phone supporting Java ME. In June 2006, Sun Company started an open source project called PhoneME, and this income became unreliable.
Of the above three branches, Java EE is obviously the one that Oracle is most interested in. But before buying Sun, Oracle Bone Inscriptions had already bought the big cow in BEA. Take the application server as an example. In China in 2008, the market share of BEA's WebLogic was 27%, second only to IBM's WebSphere(29%%), Oracle's own application server accounted for 7%, and Sun's application server (GlassFish) only accounted for 4%. Take the portal server as an example. In China in 2008, BEA's WebLogic market share was 32%, one percentage point higher than IBM's WebSphere (365 365,438+0%). Oracle's own portal server accounts for 10%, while Sun's portal server only accounts for 7%. Obviously, it is not Sun's market share in this field that makes Oracle Bone Inscriptions secretly tempted, but Sun's leading position in this field as the inventor of Java language. As you can imagine, after the merger of the two companies, Oracle will inevitably split Sun's current Java EE department into two parts-writing standards and applying for patents, continuing to write standards and applying for patents, and developing application servers will be merged into the original WebLogic department. As for Sun's own application server GlassFish, at least this brand will probably not be used again. The next generation Java EE reference implementation, called Oracle WebLogic, is obviously much louder. According to the same reasoning, Sun Company's original set of things called Java Enterprise System (JES) originally occupied a small market share, but after the merger, it clashed with Oracle's existing middleware product line and will gradually withdraw from the historical stage.
Java ME is a difficult problem. With the popularity of various high-speed wireless networks, the mobile and embedded fields are becoming the next shining enterprise-level cash machines, but Oracle Bone Inscriptions has no experience in this field. My personal opinion is that Oracle Bone Inscriptions will let this department continue to operate independently for some time, but will appoint a party secretary (perhaps newly recruited from outside) to visit and study. After the new party secretary is familiar with the business of Java ME department, he will start to adjust the department, and the adjustment range will not be great.
The fate of Java FX depends on the fate of Java ME. We know that no matter how well the desktop version of Java FX is done, you can't get a check in the desktop field. If Java FX wants to make money, it must be able to run smoothly on the mobile phone-this matter is not only as simple as improving Java FX itself, but also needs to consider the processing power of the mobile phone and the efficiency of running Java virtual machines on the mobile phone.
Both Java EE and Java ME rely heavily on Java SE. If the virtual machine at the bottom is not good enough, it is no use building a framework at the top. Sun's own Java virtual machine has many unique functions and good execution efficiency. After the merger of the two companies, it is estimated that the JRockit obtained by Oracle Bone Inscriptions from BEA will suffer. Oracle is a practical company, so the next step of Java virtual machine will focus on improving the performance of the server, and the priority of some functions only for desktop use is estimated to be lower. At this point, the development trend of Java virtual machine and GNU/Linux in recent years is basically the same.
Friends who are familiar with Java development may ask: What about NetBeans? This is a treasure that Sun Company has spent 65,438+00 years carefully cultivating. Before drawing any conclusions, let's take a look at the market share of Java integrated development environment in 2008. According to the annual survey data of programmers provided by CSDN, the domestic figures are as follows: NetBeans account for 19.3%, Sun Java Studio based on NetBeans accounts for 4.5%, Eclipse JDT accounts for 52%, and JBuilder based on Eclipse accounts for 5.5%. IBM WebSphere Studio based on Eclipse accounts for 2.8%, BEA WebLogic Workshop based on Eclipse accounts for 6.5%, IntelliJ IDEA accounts for 2.2%, and Oracle's own JDeveloper accounts for 1.2%. At first glance, NetBeans and NetBeans-based integrated development environment account for 23.8% of the market share, and it seems that Oracle has taken advantage of it in vain. Think again, NetBeans is a free product. Sun Company vigorously develops NetBeans, hoping that developers can use NetBeans to deploy their products in their own deployment environments such as application servers to sell their products and services. In the past four years, although the market share of NetBeans has increased steadily year by year, its pulling effect on other products is very limited. Sun insisted on doing it because the inertia of the past ten years was too great. If he wants to stop, he must make a big fuss both inside and outside. In other words, Sun worked hard to develop NetBeans these years, which really saved face, but lost his wallet. Larry Ellision of Oracle Bone Inscriptions would never do such a shameful thing.
MySQL database
In March, 2008, Sun bought MySQL for $654.38 billion, which was regarded by Jonathan Schwartz as "the most important merger in the history of modern software". Such an important role was not mentioned in the press release jointly issued by Oracle Bone Inscriptions and Sun Company. Many people in the industry believe that MySQL is likely to be abandoned by Oracle Bone Inscriptions because of its direct competition with Oracle Bone Inscriptions's database business. If we look at the market of database business as a whole, there is no doubt about the competition between Oracle and MySQL. However, if we further subdivide this market, the conclusion may be different.
Marten Mickos, the former CEO of MySQL, once said: "If you want to succeed in open source software, then you need services: (1) people who are willing to spend time and save money; And (2) people who are willing to spend money to save time. " Take the database market as an example. Most users of MySQL services are in the first category, and a few are in the second category. Most users of Oracle services are the second type users, and a few are the first type users. MySQL has more users, because there are always more people who are unwilling to spend money than those who are willing to spend money; The quality of users in Oracle Bone Inscriptions is higher, because those who are willing to spend money are quality customers. Orqcle is more used in medium and large enterprises, MySQL is more used in small and medium-sized enterprises, but this boundary is not strict. So there is some overlap between MySQL and Oracle, but the degree of overlap is not great. The promotion of Oracle Bone Inscriptions is top-down, and an authoritative image is established through professional advertisements. When an ordinary IT practitioner mentions Oracle, he may think that Oracle is very powerful and can do any task. The promotion of MySQL is bottom-up, cultivating users' confidence through word of mouth. When an ordinary IT practitioner mentions MySQL, he may think that MySQL is also great, "enough for such an application".
But this does not mean that Oracle Bone Inscriptions is not interested in MySQL users. From June 5438 to October 2005 10, Oracle Bone Inscriptions acquired InnoDB, which is closely related to MySQL. At that time, InnoDB provided MySQL with some technologies about things and foreign keys, which were mainly used in more complex applications and had little impact on general applications. The original intention of Oracle Bone Inscriptions's acquisition of InnoDB is to suppress MySQL by taking away the excellent features in MySQL and prevent MySQL from entering the market of medium and large applications. Unfortunately, developers generally understood Oracle Bone Inscriptions's acquisition as a fear of MySQL, and further concluded that the performance of MySQL was enough to compete with Oracle Bone Inscriptions, but it helped MySQL to make a live advertisement, which accelerated the popularization of MySQL. By the time Sun acquired MySQL in March 2008, the market share of MySQL in the database business had reached more than 25%, and it had companies like Google and Yahoo! Benchmark users like YouTube.
Now everything is logical. By acquiring Sun Company, whether you use MySQL or Oracle, you are an Oracle user. Remember, after MySQL was acquired by Sun Company, it basically operated independently and its cash flow was still positive. As long as Oracle continues to support MySQL database, it will win the support of MySQL community and provide them with better products or services at the appropriate time. MySQL will still exist as an independent product in the next two to three years. In the long run, maintaining multiple products with the same or similar functions at the same time will cause confusion to users. Therefore, MySQL will eventually need to be integrated into Oracle Bone Inscriptions's existing product line, just waiting for Oracle Bone Inscriptions to digest MySQL's existing user base.
Open source code
Through this $7.4 billion transaction, Oracle Bone Inscriptions will also receive a free gift: according to the number of lines of source code, Oracle Bone Inscriptions will become the world's largest entity contributing to the open source community. From the operating system (OpenSolaris) to the programming language (OpenJDK), from the database (MySQL) to the application server (GlassFish), from the development tools (NetBeans) to the office suite (OpenOffice), these source codes cover almost everything. This gift came suddenly, so it is estimated that Oracle Bone Inscriptions hasn't figured out how to deal with it.
Oracle Bone Inscriptions's attitude towards open source can be summed up in four words: utilitarianism. If an open source software is good enough, take it and integrate it into Oracle Bone Inscriptions's products. There is no need to pay any fees for this. In Larry Ellision's own words: "I can't write a check for hundreds of millions of dollars for open source software, because it won't give us an advantage in the competition." What we can do, others can do. "It's a pity that Oracle Bone Inscriptions acquired InnoDB, but it didn't attract MySQL users; Oracle launched Oracle Enterprise Linux, but it has been unable to attract users of Red Hat. The value of open source lies in the community formed around technology, including the developers and users of technology. Only when this community becomes bigger can it bring economic value to community leaders. At this point, Martin Mickos saw it more thoroughly than Larry Ellision.
Compared with Oracle, Sun obviously has more practical experience in open source. In the past four years, as long as it is an open source software asset, Sun Company has basically opened the source code-not only the software, but also the design of its latest version processor, UltraSparc T2. No wonder someone commented: "In this vigorous open source movement, Sun Company is undoubtedly the most thorough one. From processors to operating systems, from programming languages to development tools, from databases to office environments, there is no exception. " The problem is that, as a listed company, the open source measures failed to get the company out of financial difficulties, and were finally taken away by Oracle Bone Inscriptions. How much of Sun's "actual combat experience" can be used for reference by Oracle Bone Inscriptions is a question worthy of in-depth discussion.
label
In June 2004, I joined Sun Company with admiration. In the past four years, Sun has been in turmoil. Structural adjustment is carried out every three to five months, which makes everyone nervous for a while. Over the past few years, I have experienced many adjustments, large and small, and gradually adapted to it. In this economic crisis, I knew the company was nervous, but I didn't expect to sell it so soon. We borrow a few words from the Diamond Sutra as the conclusion of this article.
Everything has its way, like a dream, like a bubble, like dew, like electricity, and so on.
About the author:
Jiang, who received a bachelor's degree from Tsinghua University and a master's degree from the University of Illinois at Urbana-Champaign in 2000, is currently a senior manager of Sun China Technology Community and a visiting lecturer at the School of Software of Beijing Jiaotong University. Before joining Sun, Jiang worked in Beijing Zhongji Information Technology Co., Ltd. (a subsidiary of AMD), responsible for the restoration and driver development of Linux operating system. Prior to this, I worked for Jiang Navigation and Control Company in the United States, responsible for the research and development of inertial navigation, artificial intelligence and embedded systems. Jiang's personal blog address is/